Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/tomevault-io/skills-registry --skill mathprove-skill명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
| Use when this capability is needed.
> Use when this capability is needed.
Review architecture and API design for the vfs-s3 project. Use when the user mentions @architect, asks to review an issue's design, discuss module boundaries, API shape, or architectural decisions for vfs-s3. Also trigger when the user wants to create an ADR (Architecture Decision Record) or evaluate a technical approach for the project. Intended for dispatch from Codex automation or Claude routines; GitHub trigger phrase: @vfs-s3-bot please prepare design doc Use when this capability is needed.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | mathprove-skill |
| description | | Use when this capability is needed. |
本技能把证明当成可执行、可复核的软件流水线。每个结论必须绑定物理证据。
0.1 目录写入边界 — 只能在 WORKSPACE/ 下读写产物。禁止在 skill 包内写入任何临时文件。
0.2 禁止角色扮演 — 不得在主线程假装三人格。MAGI 必须通过 scripts/magi_plan.py 实现。
0.3 门控写入 — 仅当 MAGI=APPROVED + SymPy 通过 + Lean4 通过 + 证据包齐全,才允许写入 draft 并标记 passed。
0.4 终局约束 — final_audit.py 输出 APPROVED 之前,不得宣称"证明完成"。
0.5 可复核 — 每次工具调用的命令、输入、输出、退出码,必须落盘 workspace。
WORKSPACE/runs/<run_id>/
├── problem.md / assumptions.md / manifest.json / status.json
├── plan/ steps.json + steps_readme.md
├── magi/ step_XXX_vote.json
├── sympy/ step_XXX_check.py + step_XXX_out.txt
├── lean/ StepXXX.lean + step_XXX_lean.log
├── draft/ step_XXX.md + proof_draft.md
├── evidence/ step_XXX_evidence.json
├── audit/ audit.json + Solution.md + FAILURE_REPORT.md
└── logs/ init_check.log + tool_calls.log + errors.log
run_id 格式:YYYYMMDD_HHMMSS_<short_hash>
| 阶段 | 步骤 | 详细 SOP |
|---|---|---|
| A. 预检 | check_routes → 创建 run 目录 → 写 problem.md + assumptions.md | — |
| B. 规划 | magi_plan → steps.json → 结构校验 | — |
| C. 逐步循环 | 对每个 step: proposal → MAGI → SymPy → Lean4 → evidence → draft | sop/magi.md sop/sympy.md sop/lean.md sop/evidence.md |
| D. 审计 | final_audit → audit.json + Solution.md | sop/audit.md |
失败时参考 sop/error.md。
每个 step 必须原子化(一次只做一个可验证的最小推导单元),包含:
id, title, claim(含 LaTeX)inputs(前置 step id 列表)method(sympy/lean/both)sympy: { required, check_goal }lean: { required, theorem_goal }evidence_required, risks, accept_criteriaphase: init → planning → step_loop → auditing → done | failed
step.status: pending → magi_approved → sympy_passed → lean_passed → passed | failed
使用 ProofSearchTree 管理,非法转移抛 ValueError。详见 sop/runtime.md。
| 模块 | 路径 | 职责 |
|---|---|---|
| ProofSearchTree | runtime/proof_tree.py | 状态机管理 |
| ErrorClassifier | runtime/error_classifier.py | 三级错误分类 + RetryBudget |
| ParallelRunner | runtime/parallel_runner.py | N 分支并行竞速 |
| SafeVerify | runtime/safe_verify.py | Lean4 白盒审计 |
| Orchestrator | runtime/orchestrator.py | 一键编排入口 |
详细 API 参考 sop/runtime.md。
| SOP 文件 | 加载时机 |
|---|---|
sop/magi.md | 阶段 C — 每个 step 验证前 |
sop/sympy.md | 阶段 C — MAGI 通过后 SymPy 验证 |
sop/lean.md | 阶段 C — SymPy 通过后 Lean4 验证 |
sop/evidence.md | 阶段 C — 验证全部通过后写证据+草稿 |
sop/audit.md | 阶段 D — 终局审计 |
sop/error.md | 任何验证失败时 |
sop/runtime.md | 需要 Runtime API 参考时 |
Converted and distributed by TomeVault — claim your Tome and manage your conversions.